Open Semantic Standards are Key to Scaling Agentic AI Across your Enterprise

Enterprise agentic AI requires a shared semantic understanding across multiple systems and domains. Open semantic standards make that meaning portable, transparent, consistent and reusable. While semantic layers based on proprietary standards might initially seem to be quick and more convenient, they limit the ability to scale, expand and reuse enterprise knowledge across different domains. In fact, proprietary standards to tackle specific use cases are completely at odds with any AI strategy that aims for a cross-functional, cross-domain enterprise orchestration of agentic AI.

As a rule of thumb: Proprietary semantic standards in graph solutions are mainly sustainable for its vendors. For customers and their users it will become a lock-in. In an enterprise agentic AI ecosystem, this can quickly eradicate any efficiency and productivity gains companies are hoping for with AI.

To understand this, it's worthwhile to listen to Thierry Platrier, Cyber Data Senior Analyst at Airbus, at 2026's Data Week Leipzig. His most poignant statement was: "Data is no longer the scarce resource, shared understanding is. Data is cheap. Knowledge and coherence are priceless."

The keywords to focus on in his quote are "shared" and "coherence".

"Shared" because enterprises will only be able to answer the complex questions of the future and to orchestrate their agentic AI ecosystem across all departments, if all systems and applications have a shared understanding of what the data means they are working with. There will be dozens of agentic AI in each company accessing hundreds of systems and thousands of data soures, rummaging through millions of data sets. Each agentic AI will have its own logic. Each data source and system already has its own logic. Your infrastructure, data and knowledge already are siloed. A proprietary semantic layer only adds another silo to the lot. It will work in one or two systems but not the rest of your infrastructure. A semantic layer based on open standards provides the flexibility to work with any system and application, no matter the system's vendor or if it's your own team's development. Thus, it provides the framework to ensure a shared understanding of data meaning across your entire diversified infrastructure.

"Coherence" because only then will the output produced by your AI fleet be reliable, reproducible and ultimately actionable. A vendor's proprietary semantic standard can provide coherence in the vendor's system. Open semantic standards ensure that your entire ecosystem of agentic AIs can be covered.

Open semantic standards are key to your data sovereignty

Like data and its sources, ontologies and knowledge graphs are tools that must be able to adapt to the dynamic changes an enterprise faces each day. Meanings, processes, rules and data sources might change due to semantic drift. So can employed agentic AI and use cases.

In that context, digital and especially data sovereignty means your organization retains genuine control over three things: your data, your semantic model, and the economics of changing both.

Control over data requires that all data and business context knowledge are under your enterprise's own governance and not exclusively stuck inside a vendor's managed environment. It also requires that existing source systems do not need to be migrated or duplicated to be usable. Rather than moving data into a new proprietary environment, interfaces are attached to the data where it already lives.

Control over the semantic model requires transparency. The ontology needs to be understandable, not a proprietary black box inside a platform only the vendor's experts have the key to. Moreover, to retain the flexibility needed to adapt to dynamic changes, the ontology, and the underlying data model should always belong to the enterprise, not a vendor's platform. Not only does this ensure that your organization can expand its ontologies into new domains as the business evolves, combine multiple ontologies for cross-domain, cross-functional tasks, and to reuse the same ontology across different projects and systems rather than rebuilding equivalent structures again and again. The use and support of open semantic standards are the foundation for your organization's flexibility and scalability needed in today’s digital world.

Control over the cost of change is where your data sovereignty becomes a business factor. If every extension of an ontology or every reuse in cross-functional projects requires a multi-million Euro data management project, the cost of evolution scales with the number of changes your organization needs to make. That number only grows as agentic AI spreads into more domains and use cases. Genuine sovereignty means your teams have full control of their semantic layers, infrastructure and data-to-semantics mappings. This is what separates a one-off implementation from a system that can keep pace with a business that keeps changing.

Which open semantic standards exist?

One argument by the fans of proprietary standards often is that there aren't sufficient or the right open standards available. The truth is, there are open semantic standards for every aspect of building, maintaining, using and integrating ontologies and knowledge graphs in data infrastructures. Tables 1 and 2 show a selection of the most relevant open standards - mostly from the W3C - that eccenca implements and supports in its enterprise knowledge graph platform eccenca Corporate Memory.

And these standards are no newbies. Some of them have been around for quite some time. They are tried & tested, and have been improved over years of extensive implementation in knowledge and ontology projects from science to large-scale infrastructure and cross-domain use cases. With the rise of AI, new open standards have been developed over the last few years to lay the groundwork for agentic AI interoperability. The newest relevant entry is the Model Context Protocol (MCP). It was introduced by Anthropic in 2024 and handed over to the Linux Foundation's Agentic AI Foundation in December 2025 to ensure neutral governance.

Table 1: Selection of open semantic standards available for ontology and knowledge graph solutions (used and supported by eccenca Corporate Memory)

Table 2: Selection of additional open standards relevant for ontology and knowledge graph management (used and supported by eccenca Corporate Memory)

7 questions to ask your graph technology vendor

Proprietary semantic layers deepen enterprise silos while open semantic standards offer the transparency, interoperability, and reusability needed for AI to operate across your entire enterprise. Don't repeat the past mistakes of getting vendor locked-in. Prioritize open semantic standards (RDF, RDFS/OWL, SHACL, SPARQL, MCP) for new projects. The easiest and most promissing way is to start with one cross-domain use case, and then to scale to others.

When deciding on a graph and ontology solution, ask the vendor on the following categories:

1. Export: "Can you hand us the complete ontology and instance data in a standard serialization today, without a professional-services engagement?"
2. Query portability:"Will our queries run unchanged against a second engine?"
3. Mappings: "Where does the source-to-ontology mapping live, in what format, and can we read and edit it outside your UI?"
4. Constraints and rules:"Are our validation rules in SHACL, or in platform configuration? Same question for inference."
5. Third-party access: "Can an agent we build, with a client we choose, read the model and query the data without your SDK?"
6. Coverage: "Of everything we will have built in three years, what share is expressible in the standards and what share is yours?"
7. Exit: "On the day we leave, what comes with us and what do we rebuild, i.e. in terms of access rules, workflows, UI configuration, lineage?"

The answers will provide you a good idea about who will have control of your data, semantic model and cost of change in the future. The sooner these stay under your enterprise's governance, the cheaper and faster your future AI expansion becomes. It isn't just a technical decision; it's a strategic one that determines whether your AI ecosystem scales or stalls.

eccenca will be happy to answer the seven questions and additional ones before you start your journey.